Repository navigation
feat(cube-cli): Add onboarding command with step-by-step setup instructions - #12408
paveltiunov wants to merge 2 commits into
Conversation
…ctions `cube onboarding` prints the order of operations for configuring a Cube Cloud account (login, deployment, data source, data model, build, MCP), so a coding agent told to install the CLI and configure the account can follow it instead of reverse-engineering --help. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01TH2SATgmsCV5uMyXhT8HaR
|
Preview deployment for your docs. Learn more about Mintlify Previews.
💡 Tip: Enable Automations to automatically generate PRs for you. |
|
Claude finished @paveltiunov's task in 1m 58s —— View job Review: changes requested. 0 high, 2 medium, 3 low. Two problems would mislead an agent that follows the guide literally. DetailsFindings (inline)
Verified
TestingThe 3 unit tests cover the I did not build or run anything; this review is from reading the code. Review threadsThe listing returned 0 of my own threads, so there was nothing to resolve or re-affirm. All 5 findings were posted as new threads. Progress
|
…erge, GitHub path - Count as logged in only with both a URL and a token (shared Ctx::credentials), and fail on an unknown --context like every other command. - Path 4d merges the dev branch to the deploy branch and exits dev mode, so step 5 checks the branch that got the model. - Note creationMethod github for the GitHub import path; trim the Args doc. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01TH2SATgmsCV5uMyXhT8HaR
|
Claude finished @paveltiunov's task in 56s —— View job LGTM: 0 high, 0 medium, 0 low. All five fixes from the earlier review are in f10d121, and I found no new issues. Review details
Verified
Notes (not posted inline)
TestsNot run: running Review threadsThe listing returned 0 of my own threads, so none were resolved and none re-affirmed. |
Check List
Description of Changes Made
Adds
cube onboarding, which prints the order of operations for setting up a Cube Cloud account with the CLI: authenticate, create a deployment, connect the data source, add a data model (CLI upload, GitHub, dbt, or the data model Git workflow), build and validate, then mint an API token and connect the Cube MCP server.The Cube Cloud "Get Started" page (cubedevinc/cubejs-enterprise#17069) hands coding agents a prompt that only says "Install Cube CLI https://docs.cube.dev/reference/cli#installation and configure our Cube account". This command is where the agent learns the steps without reverse-engineering
--help.--api-urlorCUBE_API_URL. It skips the login step only when a token is present too, through the newCtx::credentials(), whichCtx::apialso uses. An unknown--contextfails the same way it does in every other command.merge-to-defaultandexit-dev-mode, so the build and validate step checks the deploy branch that actually got the model.creationMethod: github, matching the documented flow.--jsonreturns{ "url", "guide" }.docs-mintlify/reference/cli.mdx.Verified with
cargo build, a cleancargo clippy --all-targets, andcargo test(150 passing, 5 of them foronboarding). I also rancube onboardingwith an empty config, with onlyCUBE_API_URLset, and with an unknown--context.🤖 Generated with Claude Code
https://claude.ai/code/session_01TH2SATgmsCV5uMyXhT8HaR